Create, Manage, and Deploy AI-Enhanced Clinical Workflows with Clara Deploy SDK

Originally published at: Create, Manage, and Deploy AI-Enhanced Clinical Workflows with Clara Deploy SDK | NVIDIA Technical Blog

The medical imaging industry is undergoing a dramatic transformation driven by two technology trends: Artificial Intelligence and software-defined solutions are redefining the medical imaging workflow. Artificial Intelligence, specifically deep learning, demonstrates great potential within radiology for disease detection, localization, and classification. It has already shown it can augment humans by increasing their efficiency and effectiveness by…

This is amazing! it helps two of the biggest bottlenecks for AI in healthcare: annotation and deployment.

Thanks for the feedback. Please share any successes or failures with us. We want to learn with you.


I am playing around with Clara Train AI and MITK. Do we have any documentation that lists detailed procedure as to how to use etc. Currently, it seems to be kind of black box for me. And I am stuck in kind of installation/setup phase. I have posted in Dev forum as well. Am new to DL, hence looking forward to learn from you all

Here are my questions

1) Nvidia registry has 13 models for different organs. After setting up the annotation server with 13 models, can you let me know what happens?

2) As an end-user, let's say I have an MITK application.

3) Connection is established between MITK and AIAA server

4) Once connection is established, I will be able to see the data in MITK

5) I have unlabelled patient data in MITK now

6) I select one of those images, mark certain points based on inspection

Can you let us know what happens after this? What does these point marking in MITK do to AIAA. How will I know the label/class of the image?

It would really be helpful if we could find any extensive and easy to understand doc.


I believe it's the similar question/doubts.. please follow the video

to understand more about MITK + AIAA workflow for annotation

Hi, I've a problem running the 'ai-vnet' workflow instance with './clara-wf test'.
Since I've got the output dicom files from stage dicom-writer, I guess it launched correctly.
But I can't see any job on Clara dashboard nor any result from render server, during the whole time.
Got no error logs from dashboard container, dose I missing anything?